tool-finder

CRITICAL E004: Prompt injection detected in skill instructions.

• Potential prompt injection detected (high risk: 1.00). The skill contains an explicit instruction to "skip the ai-coord gate," which is a directive to bypass a coordination/security control and thus is a hidden/deceptive override outside the skill's stated tool-recommendation purpose.

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (medium risk: 0.30). In the Tool Finder workflow, the agent’s Step 3 “Search current authoritative sources” will require reading outsider-authored free text from sources like GitHub issues/releases and similar community-authored pages/repositories as part of evidence gathering.